home / skills / zephyrwang6 / myskill / logo-batch-generator

logo-batch-generator skill

/logo-batch-generator

This skill batch-generates 10 professional logo variants from brand description or image reference, offering 7 styles and saving results to Obsidian.

npx playbooks add skill zephyrwang6/myskill --skill logo-batch-generator

Review the files below or copy the command above to add this skill to your agents.

Files (4)
SKILL.md
5.8 KB
---
name: logo-batch-generator
description: 批量生成品牌Logo。支持两种入口:(1)用户描述品牌/用途后批量生成10个Logo变体 (2)用户上传参考图片后分析并生成Logo。提供7种主流风格选择。调用Gemini API生成,保存到Obsidian图片目录。触发词:"生成Logo"、"设计Logo"、"做个Logo"、"批量Logo"、"帮我设计一个标志"。
---

# Logo 批量生成器

根据品牌描述或参考图片,批量生成 10 个专业 Logo 变体,支持 7 种主流风格。

## 两种入口模式

### 模式 A:文字描述入口

用户描述品牌名、用途、风格偏好 → 生成 Logo。

### 模式 B:图片参考入口

用户上传参考图片 → 分析图片元素 → 询问需求 → 生成 Logo。

## 工作流程

### Phase 1: 需求收集

**模式 A(文字描述)**:
1. 收集品牌信息:品牌名(中文/英文)、行业/用途、核心理念
2. 展示 7 种风格选项(读取 references/logo-styles.md),简明介绍每种风格特点
3. 根据品牌属性推荐 1-2 种最适合的风格,说明理由
4. 等待用户选择风格

**模式 B(图片参考)**:
1. 分析用户上传的图片,提取:
   - 图形元素(形状/图标/符号)
   - 色彩方案(主色/辅色)
   - 排版风格(字体类型/排列)
   - 最接近的风格类型
   - 情感调性
2. 展示分析结果,询问用户意图:
   - "想要类似风格的 Logo?"
   - "想保留某些元素并改进?"
   - "想要完全不同方向的 Logo?"
3. 展示风格选项并推荐
4. 等待用户选择

### Phase 2: 生成提示词

选定风格后,为品牌生成 10 个 Logo 变体的提示词:

**变体策略**(读取 references/logo-styles.md 获取模板):
- 变体 1-2:主风格,不同构图(图标居中 / 图文左右排列)
- 变体 3-4:主风格,不同配色方案(浅色/深色/对比色)
- 变体 5-6:主风格,不同图标/符号变化
- 变体 7-8:相近风格交叉(如极简+线条混合)
- 变体 9-10:额外探索方向(用户可指定)

每个提示词结构:
1. 风格描述(来自选定风格的模板前缀)
2. 品牌具体描述(名称、行业、理念)
3. 构图/配色的具体变化
4. 基础指令:`Professional logo design. Pure white background. High contrast. Scalable vector-style quality. No watermark, no mockup, no background texture.`
5. 中文指令(若 Logo 含中文):`ALL text in the logo MUST use Chinese characters with clean professional Chinese typography. Technical terms and brand abbreviations may remain in English.`

展示所有 10 个提示词概要,等待用户确认:
```
## Logo 提示词概要

| # | 变体方向 | 风格 | 配色 | 简述 |
|---|---------|------|------|------|
| 1 | 主风格-构图A | 极简扁平 | 蓝色 | 几何图标居中... |
| 2 | 主风格-构图B | 极简扁平 | 蓝色 | 图文横排... |
| ... |

确认后开始生成(共 10 张 Logo)
```

### Phase 3: 批量生成

确认后逐个调用脚本生成:

```bash
python3 scripts/generate_image.py \
  --prompt "提示词" \
  --output "/Users/ugreen/Documents/obsidian/09image/MMDD-品牌名-logo/01-变体描述.png" \
  --api-key "AIzaSyDvvGGRbH4Os3Er0dYi0kE_AzE3_2b_Az8" \
  --api-url "https://generativelanguage.googleapis.com" \
  --model "gemini-3-pro-image-preview" \
  --aspect-ratio "1:1" \
  --resolution "2K"
```

**注意事项**:
- Logo 默认使用 **1:1 正方形** 比例
- 分辨率默认 **2K**
- MMDD 为当天日期
- 品牌名取简短英文/拼音(如 spaceai、mylogo)
- 文件命名:`01-风格-变体.png`(如 `01-flat-icon-center.png`)
- 每张间隔 2-3 秒避免限流
- 失败的图继续下一张,最后汇报

### Phase 4: 展示与迭代

1. 生成完成后,列出所有成功的 Logo 路径
2. 询问用户:
   - "哪几个方向最喜欢?"
   - "需要基于某个方向继续迭代吗?"
   - "需要调整配色/形状/文字吗?"
3. 若用户要求迭代,基于选中的 Logo 调整提示词重新生成

## 保存路径规范

```
/Users/ugreen/Documents/obsidian/09image/MMDD-品牌名-logo/
├── 01-flat-icon-center.png
├── 02-flat-text-horizontal.png
├── 03-flat-dark-theme.png
├── 04-flat-gradient-accent.png
├── 05-flat-alt-symbol.png
├── 06-flat-abstract-shape.png
├── 07-lineart-monoline.png
├── 08-lineart-detail.png
├── 09-explore-badge.png
└── 10-explore-gradient.png
```

## API 配置

| 配置项 | 值 |
|-------|---|
| API URL | `https://generativelanguage.googleapis.com` |
| API Key | `AIzaSyDvvGGRbH4Os3Er0dYi0kE_AzE3_2b_Az8` |
| Model | `gemini-3-pro-image-preview` |
| 默认比例 | 1:1(正方形) |
| 默认分辨率 | 2K |

## 7 种可选风格速览

| # | 风格 | 一句话特征 | 适用场景 |
|---|------|----------|---------|
| 1 | 极简扁平 | 几何+纯色,干净利落 | 科技、SaaS、互联网 |
| 2 | 线条艺术 | 单线条,优雅精致 | 工作室、设计品牌 |
| 3 | 字母标志 | 首字母创意变形 | 企业、个人品牌 |
| 4 | 徽章印章 | 圆形边框,复古仪式感 | 社区、播客、教育 |
| 5 | 吉祥物图标 | 角色化,记忆点强 | 游戏、教育、社交 |
| 6 | 渐变现代 | 丰富渐变,科技感强 | AI 产品、App、金融科技 |
| 7 | 手写书法 | 手写体,温暖有温度 | 博客、餐饮、手作 |

详细风格模板和配色参考见 references/logo-styles.md。

## 常见用法

| 用户说 | 操作 |
|-------|-----|
| "帮我设计一个 Logo" | 模式A:收集描述 → 推荐风格 → 生成 |
| "做个 Logo,科技风" | 模式A:直接推荐极简/渐变风格 |
| "参考这张图做 Logo" | 模式B:分析图片 → 询问意图 → 生成 |
| "我喜欢第3个,再做几个类似的" | 迭代:基于选中方向扩展生成 |
| "把第5个改成蓝色" | 迭代:调整配色重新生成 |

Overview

This skill batch-generates professional brand logos from either a text brief or a reference image. It produces 10 logo variants per job across seven popular styles, calls an image-generation API, and saves outputs to a defined Obsidian image folder. Triggers include common Chinese phrases mapped to the generation flow.

How this skill works

There are two entry modes: A) text description where the system collects brand name, industry, core idea and recommends styles; B) image reference where the system analyzes shapes, colors and typography then asks the user for intent. After style selection the skill builds ten structured prompts that vary composition, color and iconography, calls the Gemini image model for each prompt, and writes the PNG files into the Obsidian images directory.

When to use it

  • You need a quick set of diverse logo options from a single brief.
  • You have a reference image and want logos that keep or remix its elements.
  • You want consistent output rules (1:1 ratio, 2K, vector-like look) for branding review.
  • You plan iterative refinement: pick favorites then regenerate variations.
  • You want style-driven exploration across seven mainstream logo types.

Best practices

  • Provide a short clear brand brief: name, industry, core values and intended use.
  • Select or confirm 1–2 recommended styles to narrow the generator’s focus.
  • When uploading references, indicate which elements to keep, change, or discard.
  • Use concise English or Chinese brand names; Chinese text rule is enforced when needed.
  • Review generated set, mark favorites, and request targeted iterations instead of full regenerations.

Example use cases

  • Startup needs 10 logo concepts for investor pitch and web header.
  • A cafe uploads an old logo to preserve a mascot while modernizing colors and type.
  • A designer quickly prototypes lettermark and badge options for a client presentation.
  • An app team explores gradient vs flat interpretations for an icon set.
  • A podcast owner requests retro badge and modern line-art variants to compare tones.

FAQ

What file names and folder structure are used?

Files are saved to an Obsidian image folder named with date and short brand ID. Each file follows an indexed name like 01-flat-icon-center.png to keep variants organized.

Can I change resolution, aspect ratio or model?

Defaults are 1:1 ratio and 2K resolution using the configured Gemini model. Those parameters can be adjusted in the generation script before batch runs.